Skip to main content
summaryrefslogtreecommitdiffstats
path: root/org.eclipse.egit.ui/src/org/eclipse/egit
AgeCommit message (Collapse)AuthorFilesLines
2015-09-20Consider initial selection in Reflog view.Tobias Baumann3-6/+37
SWTCommit is adaptable to an instance of type Repository so that ReflogView can show the repository after opening. Bug: 475836 Change-Id: I7ca51f2935ab082e9a3fab04a9b4c3bae093d224 Signed-off-by: Tobias Baumann <tobbaumann@gmail.com>
2015-09-17[historyView] Added "Focus On This File" menu for selected fileAndrey Loskutov4-3/+37
Selected file in commit files list have now new "Focus On This File" context menu, which focuses history view on the selection. This is a handy option to change the history view focus and also fixes issue with older Eclipse versions which do not have Show In -> History View entry. Bug: 440994 Change-Id: Id31c410e4df24e2959c094fb188ed69a4b04e818 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-17Focus file list in GitHistoryView when context menu is shownThomas Wolf1-0/+9
Force focus to CommitFileDiffViewer to ensure that the "Show In" submenu works. Bug: 477510 Change-Id: I4faddbda611d2f9d2cbf743b8ac6545dcb1e5e51 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-16Improve performance of HyperlinkTokenScannerThomas Wolf1-12/+154
My initial implementation of this was too inefficient. It invoked the HyperlinkDetectors for every character, which effectively re-introduced bug 415537 and made editing long commit messages with only a few links a pain. Therefore, compute all hyperlinks on a line and cache them once computed. Use this cache while we're on that line, and re-compute it when the current position passes to a new line. Add an internal subclass of the standard URL hyperlink detector org.eclipse.jface.text.hyperlink.URLHyperlinkDetector because that one is only capable of finding hyperlinks at the start of the region. Others, like those from mylyn, do find all links in a region on their own. For an example commit message see: https://bugs.eclipse.org/bugs/show_bug.cgi?id=415537 Change-Id: I464f321d5f5d62836d031615b0c6252a54ee893f Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-16Add a sort order to the unstaged changesDenis Zygann5-9/+180
Add a button to sort the unstaged changes in alphabetically order or in a grouping order(sort order: Untracked, Missing, Modified), which is also alphabetically sorted. Bug: 473919 Change-Id: Ic59dd8f8d5f621dacfd06063d11a954d311a3e95 Signed-off-by: Denis Zygann <d.zygann@web.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-16Always run preferences listener UI update code in UI threadAndrey Loskutov1-3/+9
Bug: 477634 Change-Id: I5028e0e555a42683caf824e7f0366060a7d8f01d Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-16Specify modal dialog style for results of "dry run" in push dialogAndrey Loskutov5-8/+17
Bug: 391182 Change-Id: Ia86cbf6b12ae94366eb38bd1d39f2bb571e8d48b Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-15Merge branch 'stable-4.0'Matthias Sohn2-206/+161
* stable-4.0: Prepare 4.0.3-SNAPSHOT builds EGit v4.0.2.201509141540-r Don't use IPath.equals() since it is broken on Windows [Importer] More direct re-use of EasymportWizard Update build to final R20150821153341 Orbit repository for Mars.1 Fix "Import changed projects" triggered from filtered history Update com.jcraft.jsch to 0.1.53 Always supply repository to the compare input Change-Id: I4a07a16c51e578f0fd3631747a01ee2929919c46 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-13Merge changes Ief5d8000,I18808837Matthias Sohn6-28/+210
* changes: Introduce new preference page for Staging view [scalability] Staging view should handle huge amount of entries
2015-09-13Introduce new preference page for Staging viewAndrey Loskutov3-0/+61
Add new preference to configure the threshold for automatically disabling auto-expand and enabling compact tree mode if number of elements exceeds the threshold. Bug: 474064 Change-Id: Ief5d80004c03a76f046339818ee37b56fd70b5ae Signed-off-by: Andrey Loskutov <loskutov@gmx.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-13[scalability] Staging view should handle huge amount of entriesAndrey Loskutov3-28/+149
Automatically disable auto-expand and enable compact tree mode if number of elements exceeds 10000. A new preference is introduced for configuration of this value, UI for it will be added in the next commit. Bug: 474064 Change-Id: I18808837dbc26ecf98753b8cd8edd988728d574e Signed-off-by: Andrey Loskutov <loskutov@gmx.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-13Check for objects adaptable to IResource in ShowBlameActionHandlerThomas Wolf1-13/+9
Includes a new test. Some refactoring in LocalRepositoryTestCase to avoid having to duplicate test repo creation, and to reduce code duplication. Bug: 401156 Change-Id: Ib78d62f11d3a5a52c672bb192e8baba189c49829 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-12Fixed NPE in StagingView if no repository is selectedAndrey Loskutov1-12/+25
This is a follow up on https://git.eclipse.org/r/54672. If Staging view was opened without selecting any repository, changing any preference in any of the Git preference pages caused NPE in IndexDiffCache.getIndexDiffCacheEntry(Repository). Marked currentRepository field as Nullable and fixed all other possible NPE places. Enabled org.eclipse.jdt.core.compiler.problem.nullUncheckedConversion to show warnings if a value with unknown nullness is used where a NonNull value is required. Change-Id: Iab52940a67fe0205e16fa6bc5027948a4d819ff2 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-11Add check for warnings and errors before commitPawel Nowak7-51/+410
- Add optional check for warnings and errors before commit which can be configured in preferences - Allow to configure if and when a warning should be displayed and if warnings/errors should block the commit - Add a flag to the staging view which allows to override blocking ad-hoc for the next commit Bug: 475559 Change-Id: I352c33d0fed1b610babda2992c6b1b211952dfb2 Signed-off-by: Pawel Nowak <siersciotluk@gmail.com> Signed-off-by: Andrey Loskutov <loskutov@gmx.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-11Minor clean ups in new CommitMessageViewerThomas Wolf5-114/+157
* Former inner class CommitMessageViewer.ObjectLink has become global class GitCommitReference. * On JFacePreference change, it suffices to update the view; the model is not affected. (CommitMessageViewer) * Reduce code duplication in CommitInfoBuilder. Change-Id: Idf9723d385fd07d2e26319af9874b5411f5620f6 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-11Deprecate unused and obsolete methodsThomas Wolf1-2/+20
As the experience with SpellcheckableMessageArea has shown, doing syntax coloring directly in SWT is error-prone, especially in editable text viewers. These two methods are now unused within EGit and should not be used anymore. Change-Id: Ie17470740c0eaaa0ab6efd7afa581b71335199f9 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-10Rewrite CommitMessageViewer to use JFaceThomas Wolf5-242/+432
The goal is to get the commit message viewer in the git history view to use the same (configurable) hyperlink colors and other hyperlink infrastructure as the SpellcheckableMessageArea. This is complicated by two things: * custom hyperlinks to commits with custom opening, and * no clear model-view separation; getting the data from git, formatting, and coloring were all lumped together. Therefore, improve the model-view separation: * Get rid of hand-crafted SWT text range coloring. CommitInfoBuilder is restricted to getting the data from the repo, formatting, and returning semantic information (hyperlinks, header and footer ranges). It doesn't do syntax-coloring anymore. * CommitMessageViewer uses a partitioning on its document to split it logically into header, body, and footer. It also gets a custom hyperlink detector to deal with the custom commit hyperlinks. * GitHistoryPage uses a presentation reconciler with damager/repairers that know how to syntax-color each partition (and hyperlinks). * Made the HyperlinkTokenScanner a bit more flexible to rebuild the italicizing of Signed-off-by footer lines. Fixed a bug along the way (if "fill paragraphs" was on, it would put footer lines onto one line). Visually, there are no changes except the colors of hyperlinks, which now follow SpellcheckableMessageArea and are configurable. Bug: 340623 Change-Id: Ie1b5907110ba716d262e4b5773fed7e3a5150d0d Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-07Simplify hyperlink syntax coloring in SpellcheckableMessageAreaThomas Wolf3-125/+8
Turns out we don't need to use a special TextAttribute, and neither do we need a special damager/repairer. Works nicely out of the box. Change-Id: Id2f0d36f3c2018c047ad1f4cf65e148c960f236d Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-07Use standard link colors in SpellcheckableMessageAreaThomas Wolf2-18/+57
There are at least three different color definitions for hyperlinks in Eclipse: JFacePreferences, EditorsUI, and a system hyperlink color. Use standard colors instead of hard-coded blue (0,0,255). Syntax-color hyperlinks using the JFace HYPERLINK color, and make the hyperlink presenter use the configured hyperlink color for editors. Both of these can be customized by the user through (already existing) preferences and thus can be adapted to the currently used theme. (Harmonizing the discrepancy between the JFace and the EditorsUI colors is bugs 199256/135224.) Change-Id: Id0d058018ee8f6fa2279371c3f5786a3dc24b690 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-07SpellcheckableMessageArea.setText() must not bypass documentThomas Wolf1-3/+4
The sourceViewer uses an IDocument. Directly manipulating the text in the SWT widget used to display this document may bypass all kinds of JFace infrastructure, for instance hyperlinking. Definitely not good. Bug: 337288 Change-Id: Iebe30f7101540996c515d62fa9e978082cf1bd63 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-06Remove unused UItext entryThomas Wolf1-3/+0
Seems to be a leftover from commit 60765db. Change-Id: I0caebc2f27ea8652f7064ef6ae8a98d2db1731dd Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-06Use decoration color to render dirty decorator in repositories viewPeter Karena1-1/+1
Bug: 476671 Change-Id: I974b3c0e189b7a5e57a22dcc5fb5c5b079b964dc Signed-off-by: Peter Karena <peter.karena@arcor.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-06Fix sorting of nodes in repositories viewMatthias Sohn2-30/+29
RepositoriesViewSorter re-implemented sorting of nodes in a naïve way which is already implemented in RepositoryTreeNode. Instead delegate the comparison to RepositoryNode.compare(). Also improve RepositoryNode.compare() using CommonUtil comparators which ensure that sorting is case-insensitive and sort names containing numbers in natural order. This ensures all objects shown in the repositories view are sorted accordingly. This also fixes the problem that since b9dad5fa the order of repository nodes flipped when its status became dirty/clean. This was caused by RepositoriesViewSorter which compared the node label texts which changes when the dirty decorator "> " is added or removed. Bug: 476671 Change-Id: I29bfc47711ae0bd669786793ac7b40aea7fb8f55 Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-03Ensure that preferred merge strategy is stored in core preferencesLaurent Delaigue1-8/+43
Bug: 476449 Change-Id: I46817efa4ec7d82fb97e7e265a20b6ca1bd3dc64 Signed-off-by: Laurent Delaigue <laurent.delaigue@obeo.fr> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-03Make check for empty second line configurableChristian Georgi7-4/+53
Bug: 475845 Change-Id: I94146efcd7e4fdfc59af91ec11659eeb17cc8211 Signed-off-by: Christian Georgi <christian.georgi@sap.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-09-03Remove obsolete code from SpellcheckableMessageAreaThomas Wolf1-38/+0
Follow-up of commit 8a28d7ab: the SourceViewer's StyledText widget handles displaying the "hand" cursor for SWT.UNDERLINE_LINK already. Change-Id: Icdb9ba35673bb6f6fb5906d8b69ba97836753ba0 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-02Fix hyperlink display in SpellcheckableMessageAreaThomas Wolf3-38/+244
This is first and foremost a syntax coloring problem, and should be handled as such. This change implements that: * Adds a token scanner that recognizes hyperlinks using the configured hyperlink detectors. * Adds a presentation reconciler with a damager/repairer that uses this token scanner. * Gets rid of the overridden hideHyperlinks() method -- if we do proper syntax coloring, this can wreck havoc when the text is edited (the super implementation does some StyleRange cleanup; if that is omitted, a stale and possibly out-of-bounds range remains). * Uses SWT.UNDERLINE_LINK to style hyperlink partitions. * Adds a few unit tests for the tokenizing. * Adds org.objenesis to required bundles in org.eclipse.egit.ui.test to enable mocking. In my manual UI tests, this syntax-colors hyperlinks nicely in the staging view, spell-checking works, auto-wrap works, whitespace display works, the "hand" pointer appears only over hyperlinks, and bug 471355 is solved. It also has no adverse effects on the commit viewer or the git history view. Bug: 471355 Change-Id: I623ead62d9600028973353cf63183638a1e06a3e Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch>
2015-09-02Add "Compare with/Replace with" actions to ResourceMappingsStefan Dirix1-1/+23
Add the eGit "Compare with/Replace with" actions to selections which only adapt to ResourceMapping but not to IResource. This includes objects like Java-WorkingSets and custom ResourceMappings. ResourceMappings can be seen as a collection of resources forming a logical model. It often makes sense to extend the navigator view to add symbolic virtual nodes grouping the resources together. However since there is no "real" resource, the virtual node adapts only to ResourceMapping. Selecting these ResourceMappings can be seen as selecting all resources this mapping traverses. Since eGit's SelectionUtils already supports ResourceMappings (collecting contained resources), selections adapting only to ResourceMapping can be supported by simply enabling the actions on them. Bug: 474725 Signed-off-by: Stefan Dirix <sdirix@eclipsesource.com> Change-Id: I6f967f72feae10cae29505fa17b81e65e0aa87de Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-01Fixed error in decorator on bare repositoriesAndrey Loskutov1-3/+16
Change-Id: I8270c843b7a2c447b5d56f996316c93ccf39e7f8 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-09-01Don't run ResourceRefreshJob as Workspace job to avoid resource eventsAndrey Loskutov1-24/+40
ResourceRefreshJob is triggered each time user activates Eclipse window, even if there are no Git configured projects in the workspace and no changes in configured Git repositories. Unfortunately even if the job detects that it has nothing to do, it *always* triggers workspace change events and autobuild job executions. ResourceRefreshJob is a subclass of InternalWorkspaceJob, so just via entering runInWorkspace() method a sequence of calls below is always executed: workspace.prepareOperation() workspace.beginOperation() workspace.endOperation() The last one always dispatches ResourceChangeEvent to all IResourceChangeListener's and so triggers also auto build job (which usually simply doesn't do anything). Although I have no example how it can cause real harm in production environment, alone debugging the build process in Eclipse is impossible because of permanent interruptions from ResourceRefreshJob. Change-Id: Idc23fd3cf4d71d36b16af2d4855770a4b70927cd Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-30Introduce a mergeStrategy extension point.Laurent Delaigue3-1/+69
The new extension point can be used to implement and register custom merge strategies which can be used by EGit operations that require merge (merge, rebase, pull, revert, squash, stash, submodule update). Add a preference page to allow selection of the preferred merge strategy. Make sure that relevant operations use the preferred strategy. Bug: 418151 Change-Id: I4d455ce171cca13c780e72b8d27dd929b28c6b71 Signed-off-by: Laurent Delaigue <laurent.delaigue@obeo.fr> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-28Auto-configure Gerrit after ssh cloneThomas Wolf1-8/+16
ConfigureGerritAfterCloneTask.isGerrit() only worked for HTTP(S) cloning in general; for ssh cloning it worked only for git.eclipse.org. Use the same approach as for HTTP also for SSH: invoke Gerrit's "gerrit version" SSH endpoint and see what we get. Bug: 474019 Change-Id: Ic97040add0e805f34ef00f194e2a34ad3011fdcf Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-27Respect push URL's port configuration in Gerrit configurationThomas Wolf1-1/+1
If the push URI already is ssh, it most likely already has a port, and if the user cloned the repo from Gerrit, it will even be correct. Replacing that by Gerrit's default ssh port of 29418 is most confusing if that Gerrit runs its sshd on a non-standard port. Bug: 475740 Change-Id: I6043be2c101e5efd68de66bfff851b03f0ee79b2 Signed-off-by: Thomas Wolf <thomas.wolf@paranor.ch> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-24Fix "Import changed projects" triggered from filtered historyFelix Otto2-206/+161
"Import changed projects" imported too many projects when started from history view when it's filtered to some selected project. Change base class of Import Changed Projects command to fix this. With the new base class the determination of files that have been changed with a certain commit is much easier. So we can avoid the import of unwanted projects. Bug: 462977 Change-Id: Ifcfd209f33c863d93b0b7e578f83602faef2d26a Signed-off-by: Felix Otto <felix.otto@sap.com> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-23Show that there are changes in "Git Repositories" viewPeter Karena1-0/+20
Equal to the modification of a file, where the sign '>' is added to the filename, the same character is added, when a file in a repository gets changed. It's very usefull when you work with many repositories. Change-Id: I5fe30cb738c0ec5334430fc710d2d826deaebff1 Signed-off-by: Peter Karena <peter.karena@arcor.de>
2015-08-23Don't use FileNode objects as parent elements in the treeAndrey Loskutov1-1/+2
Bug: 475669 Change-Id: I1795d0145cd61f1894cc115b17794d0626ef0a55 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-20Rename usage of the label "Working Directory" with "Working Tree"Lars Vogel1-11/+11
During the review of my Git book Matthias Sohn pointed out that the term "Working Directory" is misleading and the term "Working Tree" should be used. Bug: 475537 Change-Id: I2ca1fc53160f347a56097f26c864f687b4cc207d Signed-off-by: Lars Vogel <Lars.Vogel@vogella.com>
2015-08-18Always supply repository to the compare inputAndrey Loskutov2-2/+6
Bug: 471293 Change-Id: I1b47862e261ee5ed03b840278e300cbc5cc3d24f Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-16Merge "[gitflow] Feature finish squash should allow changing the commit message"Matthias Sohn1-1/+15
2015-08-16Don't read link target for symlinks decoration to avoid exceptionsAndrey Loskutov3-21/+5
The code is not needed, because File.isDirectory() works fine for links too, and an attempt to resolve a link if the actual file is a real file will be terminated with exception. Bug: 475063 Change-Id: I23dd6989aba65987a0a657749259ed2c766b9cc8 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-16Do not iterate over all staged view entries if not neededAndrey Loskutov1-0/+6
If index has no submodules or no links, we do not need to loop over all staging view entries to set submodule or link status. Change-Id: I448e9d208dbff1653d2cf989208855cebe67152e Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-16[gitflow] Feature finish squash should allow changing the commit messageMax Hohenegger1-1/+15
- open standard commit message edit dialog to change squashed message - duplicated stripCommentLines(..) from RebaseCommand to delete comments - added UI test for feature finish squash - pulled up some redundant test code Bug: 473550 Change-Id: I028a05f21c7ca3f3a501dba160a3ed4ff035c87b Signed-off-by: Max Hohenegger <eclipse@hohenegger.eu> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-14Use default repository location for all import wizardsAndrey Loskutov8-45/+28
Bug: 395143 Change-Id: Iadc600c78e58195da4d2173764b90001fb410b2a Signed-off-by: Andrey Loskutov <loskutov@gmx.de> Signed-off-by: Matthias Sohn <matthias.sohn@sap.com>
2015-08-10Consider initial selection in Interactive Rebase viewTobias Baumann1-9/+43
If the selection, retrieved in the newly implemented init method, can be used for the setInput method, then it is used. Otherwise the selection is derived from the active editor. In both cases the existing selectionChangeListener gets reused. Bug: 473950 Change-Id: I905f3c8e54d0f22c48619246e4f1569a117ff88a Signed-off-by: Tobias Baumann <tobbaumann@gmail.com>
2015-08-02GitHistoryPage jobs should not run in parallelAndrey Loskutov2-15/+45
Diff jobs should not run in parallel to each other and to the GenerateHistoryJob. Set common scheduling rule and also added proper job cancellation on dispose() or on diff selection change. Bug: 473013 Change-Id: I9ec17c1f299e3bd726d9c8620196a92132102417 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-02Add default limit for number of shown lines in historyAndrey Loskutov7-6/+60
GitHistoryPage can show diffs up to 1000 lines very fast, up to 10000 lines reasonably fast, and starts to be unusable from 20000 lines upwards. To avoid UI freezes, a new preference is introduced in the Team->Git->History page, in the "Show" group: "Maximum lines to show for a diff", with default value of 1000 lines. If this preference is set, diff output in the History view shows only specified amount of lines. Bug: 473013 Change-Id: Ie005c179088cf94582ce730de7415a5b52caebe1 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-02Using SWT.DEFAULT halves diff widget size computation againAndrey Loskutov1-10/+1
For whatever reason, commentAndDiffComposite.computeSize() is faster if both given size hints are SWT.DEFAULT. Bug: 473013 Change-Id: I9e9dd6f17e8b2dc660d067abb65b470df8031f63 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-02Fixed double size calculation on selecting new diffAndrey Loskutov1-2/+37
Do not call resizeCommentAndDiffScrolledComposite() from controlResized() if we are already during calculation. This halves execution time. Additionally don't call both commentAndDiffComposite.setSize(size) and commentAndDiffScrolledComposite.setMinSize(size) - last one is enough for scrollable content update. Bug: 473013 Change-Id: Ie442231569fa79a7b7969e3a70448ab9d07806d3 Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-01Show user real error cause during repository search taskAndrey Loskutov1-8/+3
Properly handle exceptions (wrap to InvocationTargetException and don't "hide" to the log), show the real cause in the error message to user. Change-Id: I60b40607e1f9286f2e49e3f016a408997a47775e Signed-off-by: Andrey Loskutov <loskutov@gmx.de>
2015-08-01Decorator should properly handle bare repositoriesAndrey Loskutov4-18/+57
Bug: 474083 Change-Id: I11e7f1042af58e5752a532169fdfafe08fb06a1a Signed-off-by: Andrey Loskutov <loskutov@gmx.de>

Back to the top