Settings and activity
26 results found
-
686 votes
Jan
supported this idea
·
-
437 votesJan
supported this idea
·
-
27 votes
An error occurred while saving the comment An error occurred while saving the comment Jan
commented
I just realized that if there are many locally added files and I need to go down to 0% similarity then I receive so many matches that the actual file is not in the listing. That is I am not able to create the required search match. This makes this feature even more needed.
Jan
shared this idea
·
-
3 votes
An error occurred while saving the comment Jan
commented
I think this one can be closed because the web admin frontend is able to regain access to a server the admin has locked himself out from.
An error occurred while saving the comment Jan
commented
Hi Codice,
I'm not thinking of changing the security mode. If the server is still accessible to developers for checkout/checkin operations (but the acccess to "Repository server permissions" is blocked) the administrator would block the development team by changing the security mode.
I was thinking of something like navigating to "Repository server permissions" and in the deny dialog one could enter a security token. Something like an additional button"create security token" which triggers the server to create the token. The dialog then shows a text input field where the machine owner can enter the token from the machine. This gives access to server permissions settings and one could set the appropriate ACL.
Best regards
JanAn error occurred while saving the comment Jan
commented
I've written "The server owner can generate a security token"; of course this ability is not limited to the server owner (since it is invalid and the owner is locked out). The token can be generated by the machine owner or even by everyone, but only the machine owner can access it finally.
Jan
shared this idea
·
-
62 votes
An error occurred while saving the comment Jan
commented
Any updates on this one? This is a really important feature for me since I often have to review refactorings with hundrets of files. (e.g. method signature has changed). These are actually minor changes but with many files being changed. Back in subversion I was just reading the revision's unified diff scrolling through the changes and checking whether greater refactorings have been made here and there. This was done in 5 minutes, now I review half an hour because I have to click through each and every file!
Setting up a command line diff tool for these operations is not an option for me. I need a one click solution without the need to reconfigure my client.An error occurred while saving the comment Jan
commented
Yes, please add a feature to create a path file from a change set or even a range of change sets, or a branch. Its just a button "create patch file" from diff view and a command. :)
Jan
supported this idea
·
-
9 votesJan
shared this idea
·
-
48 votesJan
supported this idea
·
-
87 votesJan
supported this idea
·
-
35 votesJan
supported this idea
·
-
5 votes
An error occurred while saving the comment Jan
commented
Hi,
I'm afraid you mixed it up. With clicking: scrolling ok; without klicking: not ok.
I can send you a video file showing the effect by email. Whom should I send it to?
Best regards
JanJan
supported this idea
·
An error occurred while saving the comment Jan
commented
I have the same problem. @Codice: Did you klick into the code section? Once you klick into that section the scrolling is ok. But if you just move the mouse down both sections receive the scroll event.
-
10 votes
An error occurred while saving the comment Jan
commented
Hi,
although this one doesn't have that much votes yet (maybe because others don't realize the benefit :)) I'd like to know whether you plan to implement this feature in the nearer future. In my eyes this would be a big step in merge tracking capabilities.
Best regards
JanAn error occurred while saving the comment Jan
commented
Hi!
This is correct.
Best regards
JanAn error occurred while saving the comment Jan
commented
* Like "right click" on the merge link and say "delete"? That wouldn't undo the merge.
Right, that doesn't sound like the appropriate solution. You shouldn't simply delete merge links, this renders merge tracking useless.
* Maybe a combination of subtractive + delete the original merge link? (this would "corrupt" history since it would be hard to understand why the originally merged cset was created.
I agree with that the history would be corrupted in that case. You won't be able to reconstruct what was initially happening (the merge).
* Maybe mark the "merge link" as "obsolete" somehow?? - this would do it.
This sounds like the optimal solution. You could mark the merge link to have become obsolete for example with green-red dashes. The merge tracking system would consider those links as not existent. Everyone would see: "Alright, this one has been merged into branch xyz, but within that branch the resulting changeset has been undone by a subtractive merge."
This way you would be able to distinguish between the old way (mergelink not becoming obsolete) and the new implementation (merge link rendered obsolete by subtractive merge).
You might consider to add an option to the subtractive merge "with or without removing merge link"; this preserves the old functionality. However, I don't know who would want to stick with the old way.Jan
shared this idea
·
-
20 votes
An error occurred while saving the comment Jan
commented
We just had the same idea and I saw this somewhat old suggestion. So I'm resurrecting it! :)
Jan
supported this idea
·
-
12 votes
An error occurred while saving the comment Jan
commented
I've voted 3 votes on this. I agree with Göran that the update process should be unified. In my eyes, the option (3) is currently the only appropriate one for updating workspaces with local changes. But it doesn't help for branch switching.
I often find our developers to be forced to shelf and update/switch and then unshelf again because they have local changes which require a merge. This can result in huge changelists, when someone has integrated another branch (hundrets of files have changed) and just one single file matches my local changes. The result is a huge changeset although I might be introducing a few lines of code only.
I think the shelve + switch/update + unshelf workflow should be available as on-the-fly-solution in the GUI without introducing branch-local merge links. I'm thinking of rebasing my local changes onto the latest version of the repository.Jan
supported this idea
·
-
46 votesJan
supported this idea
·
-
24 votesJan
supported this idea
·
-
47 votesJan
supported this idea
·
An error occurred while saving the comment Jan
commented
* When right-clicking a changeset/line, add a choice "Annotate previous changeset" (this is so you can trace the earlier history of the line you're interested in).
This is a very important feature for us.
-
3 votesJan
shared this idea
·
-
6 votesJan
shared this idea
·
-
24 votesJan
supported this idea
·
-
13 votes
An error occurred while saving the comment Jan
commented
- Isn't cloaking applied to initial update as well? That is, the file is not being downloaded from the server, hence the application won't compile. Furthermore, what will need to be added to that list? Every developer might have different files changed locally, adjusting different components. There is simply no way to specify that kind of list.
- The files we're talking about are possibly *very* basic header files. That is, every single projects needs a recompile. In contrast, jumping through branches (bugfixing, etc.) affects only a small subset of the code.
- Shelves will solve the problem of keeping the modifications instead of rewriting them each and every time. They will not solve the rebuild problem though. Visual Studio will recognize the file being modified because the timestamp has changed.
The problem is that we have quite a big software which has a long compilation duration. Hence we need a way to switch branches without that basic files being modified.
An error occurred while saving the comment Jan
commented
Those files aren't automatically generated. E.g. the logger configuration is checked in. Likewise, you enable some debugging code in an arbitrary cpp file in the solution. Then you want to switch branch and realize that there is a checked out file which isn't supposed to be checked in. That file shouldn't be altered, but it makes it impossible for you to switch branches. Currently, you are forced to undo checkout (loosing changes, maybe shelving first, applying shelve later. Anyhow, file is altered, build is necessary). The idea is to simply change status "checked out" to "changed".
An error occurred while saving the comment Jan
commented
Hi,
I'm not suggesting to modify the behavior of the current "undo checkout" command. I am seeking for additional functionality. I'm not sure yet about how to call it. Something like "remove checkout marker" :)
UseCase: We have some files that the user changes locally very often: Global header files which add some additional debug functionality to our product; logger configuration files; xml configuration files; ini files which are versioned but are altered by our tools. Modifying those files can result in massive rebuilds (30 minutes and more). We consider to put those files on the hidden_changes list.
Suppose some of these files have been checked out, for example by the Visual Studio plugin (since they are listed in projects and one would edit them in Visual Studio). Now, it is impossible to switch branches. Furthermore, the files aren't hidden anymore.
The developer simply does not want to check in those files. But he wants to preserve his modifications and does not want the timestamp of the file to be modified (avoiding rebuilds)...Jan
shared this idea
·
Hi Pablo,
we are doing these refactors from IDE (Visual Studio) with Plugins like Resharper or VisualAssist.
Best regards
Jan