Re: [PATCH] Exit pager before calling browse-mailboxes
Hi Kevin, * Kevin J. McCarthy on Thursday, October 17, 2019 at 18:46:39 +0800: I've pushed a quick-fix commit to stable, and a better fix to master. I also cleaned up the abort/error handling when changing folder within the pager, so it returns back to the pager (except in the case of a close mailbox failure - that usually means the index will change so we need to come back up). Please let me know if you notice any problems. So far I did not notice any problems - testing master. Great, thanks! Also for the return to pager, definitely an improvement. -- LAST SHIP HOME Die Weltumsegelung der Peter von Danzig Ein Film von Michael Weber und Christian Ebert --->> https://lastshiphome.de
Re: [PATCH] Exit pager before calling browse-mailboxes
On Thu, Oct 17, 2019 at 07:42:19AM +0800, Kevin J. McCarthy wrote: I believe I know what the problem is. I was staring so hard at the wrong places I missed it. :-( I don't have time to work on it until later today, but I'll try to post a patch by tomorrow. I've pushed a quick-fix commit to stable, and a better fix to master. I also cleaned up the abort/error handling when changing folder within the pager, so it returns back to the pager (except in the case of a close mailbox failure - that usually means the index will change so we need to come back up). Please let me know if you notice any problems. Thank you, -- Kevin J. McCarthy GPG Fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA signature.asc Description: PGP signature
Re: [PATCH] Exit pager before calling browse-mailboxes
On Wed, Oct 16, 2019 at 05:32:46PM +0100, Christian Ebert wrote: I'll try. As I have got other stuff on my plate and am back into blind guessing territory - there is a reason why I wasn't more precise in my commit message(s) regarding the cause - this may take a while. Sorry. I believe I know what the problem is. I was staring so hard at the wrong places I missed it. :-( I don't have time to work on it until later today, but I'll try to post a patch by tomorrow. -- Kevin J. McCarthy GPG Fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA signature.asc Description: PGP signature
Re: [PATCH] Exit pager before calling browse-mailboxes
* Kevin J. McCarthy on Wednesday, October 16, 2019 at 09:58:38 +0800: On Tue, Oct 15, 2019 at 03:32:01PM +0100, Christian Ebert wrote: I have the following in my muttrc: folder-hook somefoldernamepattern ' push ~r>1m' This example isn't triggering the problem for me (I have subsituted '.' for the matching regexec). Would you mind double checking that it isn't a different folder hook, or something else. Indeed, it must be something else. I was able to repro it by going into a different target folder. If you could send a minimal muttrc that triggers it, that would be helpful. I'll try. As I have got other stuff on my plate and am back into blind guessing territory - there is a reason why I wasn't more precise in my commit message(s) regarding the cause - this may take a while. Sorry. -- LAST SHIP HOME Die Weltumsegelung der Peter von Danzig Ein Film von Michael Weber und Christian Ebert --->> https://lastshiphome.de
Re: [PATCH] Exit pager before calling browse-mailboxes
On Tue, Oct 15, 2019 at 03:32:01PM +0100, Christian Ebert wrote: The goal is that 'y' should _not_ behave differently. I.e., I should be able to browse-mailboxes, choose a different mailbox and open it without error. Hi Christian, Sorry for the barrage of questions in my last email. I should have actually asked questions last year instead of just accepting the patch. I agree that it shouldn't behave differently, so I will try to track down the difference and see if I can fix it "properly". There *is* a small difference between the two states of invoking , but I currently can't see how it is causing the problem. I have the following in my muttrc: folder-hook somefoldernamepattern ' push ~r>1m' This example isn't triggering the problem for me (I have subsituted '.' for the matching regexec). Would you mind double checking that it isn't a different folder hook, or something else. If you could send a minimal muttrc that triggers it, that would be helpful. Thank you, -- Kevin J. McCarthy GPG Fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA signature.asc Description: PGP signature
Re: [PATCH] Exit pager before calling browse-mailboxes
Hi Kevin, * Kevin J. McCarthy on Tuesday, October 15, 2019 at 09:09:12 +0800: On Mon, Oct 14, 2019 at 04:02:52PM +0100, Christian Ebert wrote: * Christian Ebert on Monday, October 14, 2019 at 15:50:01 +0100: +macro pager y "" "show incoming mailboxes list" As there's now a clean command instead of macro hacks - thank you, Kevin -, it may be worth considering to integrate this directly in browse-mailboxes? I forgot about that patch from you last year. However, I'd like to review the reasons for this before applying your new patch. How is this any difference from calling or from the pager? Why should 'y' behave differently? The goal is that 'y' should _not_ behave differently. I.e., I should be able to browse-mailboxes, choose a different mailbox and open it without error. What exactly are you doing that generates the error when invoking 'y' from the pager. I have the following in my muttrc: folder-hook somefoldernamepattern ' push ~r>1m' 1. I am in a mailbox, let's call it 'currentbox'. 2. I am in its *index* and press 'y' for . 3. From the browser I choose a mailbox which matches 'somefolderpattern'. 4. I open the target mailbox without errors. From the *pager* 1., 2., 3. behave the same, when i do 4. mutt barks: 'Not available in this menu'. I want mutt to behave the same as if 1. and 2. would happen in the index. Is there any reason why you shouldn't create the macro yourself in that case? 1. unexpected behaviour, even error 2. hard to find out the reason, and hard to find the cure [*] [*] I don't really understand why the error is triggered by a hook of the target mailbox combined with the fact that it is not triggered when I start the procedure in the index. It's as if there is some remnant pager state when I'm in the browser: delete-pattern indeed is only available from the index, neither from the browser (but if that was the problem, then the error would also be thrown if I start from index) nor from the pager. So, in layman's terms: For consistent and error free behavior I have to exit the 'lingering pager state'. Admittedly it feels like doctoring at the symptom, but so far it works for me[tm]. -- LAST SHIP HOME Die Weltumsegelung der Peter von Danzig Ein Film von Michael Weber und Christian Ebert --->> https://lastshiphome.de
Re: [PATCH] Exit pager before calling browse-mailboxes
On Mon, Oct 14, 2019 at 04:02:52PM +0100, Christian Ebert wrote: * Christian Ebert on Monday, October 14, 2019 at 15:50:01 +0100: +macro pager y "" "show incoming mailboxes list" As there's now a clean command instead of macro hacks - thank you, Kevin -, it may be worth considering to integrate this directly in browse-mailboxes? Hi Christian, I forgot about that patch from you last year. However, I'd like to review the reasons for this before applying your new patch. How is this any difference from calling or from the pager? Why should 'y' behave differently? What exactly are you doing that generates the error when invoking 'y' from the pager. Is there any reason why you shouldn't create the macro yourself in that case? -- Kevin J. McCarthy GPG Fingerprint: 8975 A9B3 3AA3 7910 385C 5308 ADEF 7684 8031 6BDA signature.asc Description: PGP signature
Re: [PATCH] Exit pager before calling browse-mailboxes
* Christian Ebert on Monday, October 14, 2019 at 15:50:01 +0100: +macro pager y "" "show incoming mailboxes list" As there's now a clean command instead of macro hacks - thank you, Kevin -, it may be worth considering to integrate this directly in browse-mailboxes? -- LAST SHIP HOME Die Weltumsegelung der Peter von Danzig Ein Film von Michael Weber und Christian Ebert --->> https://lastshiphome.de