Bugzilla – Full Text Bug Listing
| Summary: | Strange chars when file comparing SVN to local files | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Technology] Subversive | Reporter: | Andrew Scott <andrews> | ||||
| Component: | Connectors | Assignee: | Igor Burilo <igor.burilo> | ||||
| Status: | CLOSED MOVED | QA Contact: | |||||
| Severity: | minor | ||||||
| Priority: | P3 | CC: | a.gurov | ||||
| Version: | unspecified | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | PC | ||||||
| OS: | Windows 7 | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Andrew Scott
Those are UTF-8 encoding file marks that were inserted by notepad.exe or some similar tool. Looks like Eclipse IDE compare editor just shows those marks even when file encoding is UTF-8. Personally I think it is very good behaviour, because many programming tools fails compilation or another activities when those invisible marking bytes are present. Also, probably compare editor behaviour regarding those marks could be altered by changing Window->Preferences->General->Compare/Patch->Ignore white space option. (In reply to comment #1) > Those are UTF-8 encoding file marks that were inserted by notepad.exe or some > similar tool. Looks like Eclipse IDE compare editor just shows those marks even > when file encoding is UTF-8. > Personally I think it is very good behaviour, because many programming tools > fails compilation or another activities when those invisible marking bytes are > present. > Also, probably compare editor behaviour regarding those marks could be altered > by changing Window->Preferences->General->Compare/Patch->Ignore white space > option. How is this good? If my team is using Eclipse and Subversive then this is a problem from time to time, if we all use Eclipse/Smart SVN this is never a problem. So if SmartSVN never sees this as a probblem, but Eclipse and Subversive does then there is a problem and this is not good. It is a bug. The point here is that if people are using anything outside of eclipse and SmartSVN there is no problem, and this would be not considered a bug. However Eclipse when saving this file sometimes does something that is showing this in the file compare as a problem. And it is only Eclipse that is havong this problem, if we use any other editor that is UTF-8 compliant and save it as UTF-8 we sometimes see this in the file compare. Yet if I do not merge this, and do the same compare in SmartSVN there is no problem. Figure that out, there is abug here and it is causing us issues do not close this again unless you know what we are using. We are not using NotePad++, and this is an Eclipse bug. We have done enough tests to show this to be an Eclipse bug, if you require more info then please ask, do not assume we are usingf NotePad++, do not assume that anything outside of what I have describe is causing this. Everyone is using Eclipse here, but the comparison in my case is eclipse, where the rest of the team use SmartSVN. SmartSVN does not complain about this in the file comparison window, where as Eclipse does. THIS IS A BUG. Could you at least please provide some information regarding the issue. May be a sample file in attachement? (Note: I don't need complete file content, so you may drop parts with commercial secrets.) I'm asking you for sample data because I can't be sure if the lines you posted here corresponds to realty or not after going through bugzilla database and then bugzilla Web-interface. I mean those symbols are binary mark and representing it through text interface is no good. I tried to reproduce the situation once again but with no success. So, for now I'll be waiting for the sample file(s) and steps to reproduce. I will see what I can do, the problem is that it is a very intermittent one. And we haven't seen it for awhile now. But as soon as I see it again it will be attached here. The problem I have is that I have no idea what is actually causing this, the common element is Eclipse and I say this because it is only Eclipse and the file comaparison window that we ever see these characters. If I use any other SVN comparison tool, I never see this in those applications. Just every now and then it pops up in Eclipse, it is almost like the UTF-8 headers are getting corrupt or it is not reading it right and/or getting confused with it under certain circumstances. Most use SmartSVN and never see this problem, and I am the only one who uses Eclipse and Subversive to do a file comparison where it eventually gets seen. But as stated this is a rare occurance, but it does happen. Created attachment 185291 [details]
UTF-8 problems in this file
Ok I was able to get it to happen again.
All we do is change the extension of the file so UTF-8 is switched off then remove the extra chars that are at the top of the file.
The problem is that I am still not 100% convinced this is an Eclipse issue, I have also reported this with Adobe as the editor is the plugin to Eclipse that just might be the issue.
But I wanted to report it just in case it isn't.
(In reply to comment #8) The file is really broken, since the UTF-8 mark is present twice (while it should be present no more than once). The only parts of Subversive that works directly with the file content are external client libraries (in time of checkout, update or modifying svn keywords). So, in a long time I've tried to reproduce this behaviour with a set of different connectors, but without any success. I suppose there are no issue like this in those client libraries, so I've nothing to report to the library developers and there is nothing I can do on the Subversive side. Since you've asked me not to close the report I won't for awhile, but at least I'll lower the report importance since it seems most likely related to some other Eclipse plug-ins (editors, for example). This issue has been migrated to https://gitlab.eclipse.org/eclipse/subversive/subversive/-/issues/68. |