Closed Bug 98296 Opened 23 years ago Closed 22 years ago

helper app defined from prefs panel won't work

Categories

(SeaMonkey :: Preferences, defect, P2)

PowerPC
Mac System 9.x

Tracking

(Not tracked)

VERIFIED WORKSFORME
Future

People

(Reporter: bugzilla, Assigned: samir_bugzilla)

References

Details

(4 keywords, Whiteboard: [correctness])

found using 2001.09.04.11-comm bits on Mac OS 9.1. if i define a helper app from the preferences panel for a mimetype where there's no OS-defined helper app, the helper app dialog somehow does not pick it up. oddly enough, this is not a problem if i define the helper app directly from the helper app dialog via the Advanced button. 1. go to the Helper Applications panel in the prefs dialog. 2. click New Type, and fill out the resulting dialog. eg, i used the following: Description: Portable doc format extension: pdf mimetype: application/pdf application: [path to Adobe Acrobat Reader] 3. clicked OK, OK. 4. clicked on a pdf link, eg, one of the links at http://www.mozilla.org/quality/browser/front-end/testcases/plugins/acrobattest.html 5. looking at the resulting helper app dialog. expected: the helper app dialog should have "Open using Acrobat Reader" selected. actual: the helper app dialog opens with "Save this file to disk" selected. moreover, the Open option has "Open using <no application specified>".
nominating. not a problem on winnt or linux.
marking nsbranch+, pdt. How close are we to fixing? What does the fix involve?
Keywords: nsbranchnsbranch+
Blocks: 99227
Priority: -- → P2
Target Milestone: --- → mozilla0.9.6
->095 since we're asking for this in the current release. Is this going to happen in the next couple days?
Keywords: mozilla0.9.4
Target Milestone: mozilla0.9.6 → mozilla0.9.5
Accepting.
Status: NEW → ASSIGNED
se, can you help me by isolating between which two builds this first started occuring? (If sweetlou doesn't have enough nightlies ftp.mozilla.org seems to be a good place to turn to.) Thanks.
I cannot reproduce this with a build pulled a couple of days ago from the trunk. I get the expected results as listed by se in the original report. If I add the MIME type and click on a pdf link from the cited test page, I get the dialog with the "Open using Acrobat Reader" radio button selected. If I then proceed to remove this MIME type association via the Helper Apps pref panel and click on a pdf link, I get the the dialog with "Save this file to disk" selected and the open option has the "Open using <no application specified>" radio button selected. se, is this branch-specific?
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → WORKSFORME
Adding correctness Status Whiteboard, correct/expected behavior does not occur.
Whiteboard: [correctness]
fooey --this is still a problem using either commercial branch [2001.09.24.03-0.9.4] or trunk [2001.09.24.13-trunk] bits on Mac OS 9.1. :( reopening. samir, perhaps it works for you because application/pdf might already be predefined for you mac [check your Internet Config]. this bug, however, seems to occur with any user-defined mimetype [not already defined by OS] on mac --so, try testing with another type. eg, use the following when defining the handler in the prefs panel: extension: exe mimetype: application/octet-stream application to use: [i arbitrarily selected the path to bbedit v5.1] sample file to click: go to ftp://ftp.mozilla.org/pub/mozilla/nightly/ and go into a build dir that contains an .exe file, then click it.
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
We'd like to take all regression fixes when available. - PDT Samir - Pls look into this one, and give us an ETA.
Whiteboard: [correctness] → [correctness],PDT
Samir, what's your progress on this bug?
Can reproduce given se's new set of steps. Still would like the two builds between which this regression cropped up. Initial investigation points to an RDF problem when two clients are sharing a datasource. I think the RDF implementation should handle synchronization rather than have multiple clients follow some protocol.
Status: REOPENED → ASSIGNED
OK, on a new profile I can't reproduce this even using se's new steps. And I no longer have an old profile on my debug machine. se, Can you confirm that this doesn't affect new profiles?
Ignore my last comment. This bug only affects users in the same browser session that the MIME-type was created in. If they restart their browsers the mapping takes effect.
samir: i restarted the browser, and the user-defined type *still* didn't take affect. tested using 2001.09.25.03-branch comm bits on OS 9.1 [emul over os x].
drat, i tried with a new profile, and the user-defined pref still doesn't go into effect after a restart.
hokay, samir and i have narrowed this down to occurring btwn 2001.08.01 and 2001.08.06 trunk [well, pre-0.9.4 branch, anyhow!]. ie, not a prob with 2001.08.01 or earlier bits. another observation: when going thru these older builds, i found that samir's workaround to restart the app does work. so, a difference btwn trunk and branch: branch bits don't offer a restart workaround. alas. :(
samir, i can actually repro this using 2001.08.01.04 [trunk] bits. i'm gonna look at mozilla builds btwn 7/26 and 8/1...
scratch that --cannot find mozilla bits btwn 7/26 and 8/1. so, it looks like something was checked in btwn the 2001.07.26.21-0.9.2 [albeit branch] and the 2001.08.01.04 [albeit trunk] builds that caused this regression.
This needs an ETA date in the status whiteboard today.
Whiteboard: [correctness],PDT → [correctness], PDT, ETA: 09-27
Now I simply cannot reproduce this under any conditions on my machine. I even tried by serving up a new unused MIME-type (configured my server to serve application/x-samir for files with .sbg extensions) and this bug simply does not occur in my setup for ns 2001-08-01-04-trunk, ns 2001-08-06-04-trunk, moz 2001-09-21-03-trunk, moz 2001-09-26-03-0.9.4, or moz 2001-09-18 developer debug builds. Have a couple of ideas to experiment when se's machine becomes free. se, have you been able to reproduce this on any other machine consistently?
Whiteboard: [correctness], PDT, ETA: 09-27 → [correctness], PDT, ETA: n/a
BTW, even the .exe testcase works for me.
Can't reproduce on pchen's debug mozilla build from this morning. On Jimmy's G4 I reproduced this once but could not ever reproduce it with the exact same steps again. The problem where the MIME-type mappings don't get deleted from the mimeTypes.rdf datasource even if they are removed from the helper apps prefs panel may have something to do with it: as suggested by law, the uriloader may get a different reference to the mimeTypes.rdf datasource than the one used by the prefs panel. In addition, the two times I have actually witnessed the problem have only been for the aplpication/octet-stream and exe combination -- one which is unlikely to be used by mac users. Although, I admit that the problem can possibly be wider than just that combination.
OK, I saw this (only once) while I was debugging and the problem was that RDFXMLDataSourceImpl::Init(const char* uri) was failing. Since then I haven't been able to reproduce this while debugging and hence don't know what was failing exactly. Any ideas welcome.
nsbranch-, not widely reproducible, no fix, and too risky to change these areas at this point.
Keywords: nsbranch+nsbranch-
Target Milestone: mozilla0.9.5 → mozilla0.9.7
Removing [PDT] grafitti because this was minused by component team.
Whiteboard: [correctness], PDT, ETA: n/a → [correctness]
Blocks: 107067
Keywords: nsbranch-
Moving to mozilla0.9.8.
Target Milestone: mozilla0.9.7 → mozilla0.9.8
Target Milestone: mozilla0.9.8 → mozilla0.9.9
tentatively nominating for nsbeta1. argle, i was able to reproduce this using a new profile on mac 10.1.2 with 2002.01.25.03 comm trunk bits. however, there were a few times where i *wasn't* able to repro --but that was with an existing profile.
Keywords: nsbeta1
Prefs triage team: need a reproducible testcase.
Target Milestone: mozilla0.9.9 → Future
No longer blocks: 107067
using 2002.06.27.05 branch comm bits on mac 9.1 and 10.1.5, i'm unable to repro this so far --helper apps that i define from the prefs panel now work. the application is noted in the dlg, and launchable. marking wfm --but do reopen if anyone else sees this with recent builds (branch or trunk).
Status: ASSIGNED → RESOLVED
Closed: 23 years ago22 years ago
Resolution: --- → WORKSFORME
v.
Status: RESOLVED → VERIFIED
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.