Menu
▾
▴
#2259 Hairpin still hits barline at line break
Originally created by: *anonymous
Originally created by: pkx1...@gmail.com
Originally owned by: mts...@gmail.com
Originally reported (and supposedly fixed in issue 2130).
Pasting comment here for new issue
--snip--
Comment 8 by przemoc, Today (7 minutes ago)
Sorry for bothering you again, but original example lacked broken hairpins (that I mentioned in one of my comments in issue 2057 ) and they're still touching the span bar in 2.15.26. Moreover, changing broken-bound-padding has no effect.
--snipp
Originally posted by: k-ohara5...@oco.net
Dynamics text hitting bar lines is issue 621.
Mergedinto: 621
Status: Duplicate
Originally posted by: prze...@gmail.com
Below is simple snippet showing actually two problems:
- lack of broken bound padding for hairpin in Dynamics
- lack of broken bound padding for hairpin in Staff if it's below the span bar
As I said already a few times before, padding should be always there, but obviously in case of broken hairpins this padding should be very small (like the one you can see if you add f\> g b\! decrescendo to the right hand). If you think otherwise, then please at least create setting that would allow changing the behavior.
\version "2.15.26"
#(set-default-paper-size "a6" 'portrait)
global = { \key c \major \time 4/4 }
rightHand = \relative c' { \clef treble c1 d e f g b \bar "|." }
leftHand = \relative c { \clef bass c1 d e f\> g b\! }
dynamics = { s1\p\< s1 \break s1 s1\f \break }
\score {
\new PianoStaff = "piano" <<
\new Staff = "up" << \global \rightHand >>
\new Dynamics = "dynamics" \dynamics
\new Staff = "down" << \global \leftHand >>
>>
\layout { }
}
Originally posted by: m...@mikesolomon.org
I can have a look at this in a couple weeks - pester me if I forget!
Originally posted by: prze...@gmail.com
To make it even more crystal clear for the second problem, "\override Hairpin #'broken-bound-padding = #10" in leftHand obviously doesn't work, but if you add mentioned decrescendo to the rightHand, then it magically starts working...
And I really don't understand why it got merged into issue 621, which is a completely different thing, right?
Originally posted by: k-ohara5...@oco.net
(No comment was entered for this change.)
Mergedinto:
Status: Accepted
Originally posted by: m...@mikesolomon.org
It's because broken-bound-padding only kicks in if there is a collision with span bars. Here, there is no collision, so it doesn't kick in. Can you try normal padding and see if that does anything?
Originally posted by: prze...@gmail.com
Thank you for your interest, Mike.
So broken-bound-padding has bad definition for my taste, because it should work always (or if you feel strongly that this is unacceptable for some strange reason -- it should work at least if there is any span bar, even not colliding one [like above]).
Going with #'bound-padding = #10 shortens only the part after break. Sorry I'm not yet proficient in LilyPond, so I'm not sure what normal padding you meant here.
Also if you see or have an idea for some workaround (especially for the first problem, i.e. hairpin in Dynamics), I'll gladly test it.
Originally posted by: prze...@gmail.com
I'm attaching another image to show how my snippet (comment 3) should be engraved.
And any of these (de)crescendos shouldn't move even by 1/10 pt if another one is removed. (Saying it just in case, because previous hairpin issue showed, that LilyPond devs apparently like to needlessly create strange corner cases, that solve problems only partially).
Originally posted by: k-ohara5...@oco.net
The intention was for all Hairpins to get either bound-padding or broken-bound-padding as appropriate, but I missed a case.
Mike wants broken Hairpins change based on concurrent hairpins for his own reasons, and wrote the code. The default 'broken-bound-padding is a function that uses Mike's code. Some of us don't want the complicated behavior, so we can
\override Score.Hairpin #'broken-bound-padding = #0.5
http://codereview.appspot.com/5587046
Labels: patch-new
Status: Started
Originally posted by: prze...@gmail.com
It's counter-intuitive then, because broken-bound-padding's name suggests it should kick in always whenever padding is applied for broken hairpin. If it's not that and there won't be any will to fix it, then I think it should be renamed (broken-collision-bound-padding?).
If padding would be always there (independently of span bar vs bar existence, whether there is collision or not), then maybe would be no need for special behavior of concurrent hairpins (that spanned across same notes should have obviously same horizontal position and length).
IMO defaults should be the most universal settings and at the same time giving same look for similar use-cases (like hairpin below span bar vs one potentially colliding with span bar). Good clean look is possibly top priority for LilyPond, even if it makes code more complex, but if you'll remember my previous sentence, I think that complexity will be be actually lowered.
Originally posted by: k-ohara5...@oco.net
After the patch, 'broken-bound-padding _does_ always take effect when Hairpins are broken across a line.
So, if we override 'broken-bound-padding = 0.5 all Hairpins will stop half a staff-space before the bar.
In LilyPond, however, properties like 'broken-bound-padding may call functions to do complicated things. If the user does not override it, 'broken-bound-padding uses a function to inspect whether it needs to avoid any span bar.
Originally posted by: lilypond...@gmail.com
Patchy the autobot says: LGTM.
Labels: Patch-review
Originally posted by: ColinPKC...@gmail.com
(No comment was entered for this change.)
Labels: -Patch-review Patch-countdown
Originally posted by: ColinPKC...@gmail.com
Counted down to 20120131, please push
Labels: -Patch-countdown Patch-push
Originally posted by: dak@gnu.org
I am making this "owner=Mike" since apparently he wrote the patch and is likely most familiar with its interactions with other patches from him in the queue at the same time. I am not sure of possible interdependencies, so it is probably best if Mike pushed this himself if appropriate.
Owner: mts...@gmail.com
Originally posted by: k-ohara5...@oco.net
commit 4dd2a57
No hairpins touch barlines
Broken hairpins that /could/ go under a bar line get the special behavior depending on concurrent hairpins, but now we can turn it off and get simple consistent behavior with
\override Score.Hairpin #'broken-bound-padding = #0.5
Labels: -Patch-push Fixed_2_15_29
Originally posted by: prze...@gmail.com
Thank you for the fixes! I'll gladly test hairpins again soon after 2.15.29 release. Sorry for being so pushy about them (and not fully understanding the patches alone), but it's all about making LilyPond better.
Originally posted by: dak@gnu.org
If "it's all about making LilyPond better", the proper way is to deliver prompt and complete feedback and analysis and make it as easy and enjoyable for developers to make a fix. This may include offering a bounty. The reporter should avoid creating the impression that the purported importance of the fix is inversely proportional to the amount of resources he is willing to invest himself.
That being said: I have seen decidedly worse.
Originally posted by: prze...@gmail.com
Sorry for disappointing you then. LilyPond is not my only concern, because then I would definitely try to understand its code and make patches myself (it may happen anyway some day...). But if someone is into it already, then such fixes possibly shouldn't be a big problem. I had times when my feedback wasn't prompt (as in case of issue 2232), but I think I was more commenting (issues 2057, 2130, 2165 and now 2259) than you can see in most other issues. Also my initial report was using piano centered dynamics already, but it got into tracker w/o it (submitter apparently thought it's irrelevant). I agree though, that I did not cover all possible variations of hairpins back then (but I commented later about most of them). I didn't thought that broadly, but developer possibly should, because (s)he is (at least should be) much more aware of implementation peculiarities and its side effects.
All that aside, "inversely proportional"? What the heck are you talking about and where it comes from? Because I don't build lilypond myself as soon as patch is in the repo to check its result immediately? Or it's because I don't offer a bounty?
I appreciate all your efforts, David, really. But turning everything into a bounty thing isn't that good (maybe you're too much into making your profitable business based on LilyPond?). And I would maybe even consider donating some general bounty, because I think LilyPond is great, but when you mention it that explicitly, I'm somehow becoming more than reluctant. Why there isn't any donate button on the homepage in the first place? PayPal, AlertPay, Flattr, plain IBAN, anything. Fix that and LilyPond dev crew will get some donations. Maybe not as much as anyone would like it to be, but surely better than nothing. Unless you think that 10$/10€ or such on the street is not even worth bending forward, because then, well, it's still like nothing I guess.
That being said: I'll try to be a better bug reporter, so don't hesitate to poke me whenever my feedbacks aren't as complete and elaborate as they should (because, you know, user often just don't know that you, the dev, is lacking necessary information in provided report for instance). Remember that sometimes casual reply/question directed to reporter (like "przemoc, mind that centered dynamics work different that in staff, so does the image below looks as it should?") can do wonders. Finally CCing reporters when adding issue to the tracker also should be done by default, yet so far it happened only once (and I have to hunt issues related to my reports myself and star them to get any updates)!
Please keep making LilyPond better. Cheers!
Originally posted by: k-ohara5...@oco.net
(No comment was entered for this change.)
Status: Fixed
Originally posted by: dak@gnu.org
LilyPond is not your only concern. But you expect it should be the only concern of developers. You feel it is shameful of them to mention "bounties" or any other kind of compensation. You make a big fuss that you poor guy "have to hunt issues related to my reports myself and star them to get any updates". You pontificate things like "Remember that sometimes casual reply/question directed to reporter (like "przemoc, mind that centered dynamics work different that in staff, so does the image below looks as it should?") can do wonders" and finish with "Please keep making LilyPond better. Cheers!".
Self-centered demanding and demeaning crap like that is what made Graham decide to finally quit working on LilyPond. I certainly would prefer if users with that "benign slave-owner" attitude just went away instead of destroying what remains of a working community.
This does not belong in an issue report, but I would hardly be doing my fellow developers a favor by letting it stand uncontradicted.
If you think the most important thing you can teach a developer is manners, do everyone else a favor and choose a drunken bully in the next pub as your teaching object instead. You'll be doing much less damage, have a better chance of actually having somebody learn something, and fewer people will be paying the price.
Originally posted by: prze...@gmail.com
Manners? I was writing at most about some good habits, that make cooperating easier. Surely they're not mandatory ones.
I can assure you I'm not self-centered and demanding one. Actually I did not demand anything from LilyPond dev team even once; I was just reporting problems (first report wasn't on bug ML, as I wasn't sure it is a bug, being new to LilyPond), bugs, stating how I think things should look engraved (sometimes firmly, I know) and asking for your fixes (when you'll be able to do them, of course) or workarounds (if possible). You're mistaking things.
Moreover, it's not shameful to mention bounties (that's why I think donate button should be on the home page, really!), but how you do it and in what context it appears (i.e. what is the real message you want to convey) is relevant (and the message in your case looked simply like performing an attack on me, really don't know why, even using 3rd person as disguise; but it was w/o good reason IMO as I was rather active in comments, yet you were jeeringly praising me for not being the worst).
But the most important thing here is that I've never said, written or even thought that "[LilyPond] should be the only concern of [its] developers", because I understand well that you all are doing it in your free time. Slave-owner attitude? Hell no. Benign meant as kind? Yes, because I don't see why I should be mean just because I've spotted some bugs or whatever. Errare humanum est.
EOT (and sorry for OT here, because it really does not belong to an issue report, but I had to respond, as I couldn't stand inexplicable ill will towards me)
P.S. I accidentally deleted first comment due to inconvenient touch pad.
Originally posted by: ColinPKC...@gmail.com
The decrescendo under the second system still extends underneath the barline, but moves away from it with the \override Score.Hairpin #'broken-bound-padding = #0.5 from comment 17.
Status: Verified