Forgot to include pd-list and l2ork-dev list... ---------- Forwarded message ---------- From: Ivica Bukvic <[email protected]> Date: Mon, Nov 2, 2015 at 12:01 AM Subject: Re: [PD] What is the purpose of iemgui_raute2dollar To: Miller Puckette <[email protected]>
So, I am about to commit a 'fix' in pd-l2ork where one can now have labels (for gatoms and arrays as well) where escaped $ is converted into ASCII 0x01 (SOH). This seems to side-step the issue for the time-being. Does this affect other objects? The ideal fix would be using negative ascii values (e.g. -0x80) but those don't seem to cut the mustard once they are received by tcl/tk... Best, Ico On Sun, Nov 1, 2015 at 11:04 PM, Miller Puckette <[email protected]> wrote: > This does need fixing... it's a primitive escaping mechanism which needs to > be replaced with a "correct" one. Unfortunately there are various places > in the code that escape stuff in various ways (to get strings past > the TCL parser or the Pd parser) and it's never been thought about in a > unified way. > > cheers > Miller > > On Sun, Nov 01, 2015 at 10:57:42PM -0500, Ivica Bukvic wrote: > > Right, but this function does not do that, it explicitly replaces # with > $. > > What does # have to do with anything? Are you saying # is being > substituted > > for $ to preserve it? Even so, this still makes no sense because if you > > create an iemgui object (pick one), then add a label with only a "#" in > it. > > Once you click apply, you get $ on the label instead. Reopening > properties > > provides you with a label that is equal to $, so this sounds like a bug > to > > me that does some sort of one-way mangling that needs to be fixed and > more > > so, it seems to me there should be a different ascii character used to > > preserve $, one that does not exist in the regular text-based ascii > > chars... e.g. in pd-l2ork we use \t to substitute for \n and thus > preserve > > endlines in comments and other places. > > > > On Sun, Nov 1, 2015 at 10:33 PM, Jonathan Wilkes <[email protected]> > wrote: > > > > > Because Pd's parser is likely to eat up a "$" followed by a number as a > > > banking > > > fee to pay the cost of sending messages. :) > > > > > > Send this from the GUI: > > > (canvas_addy) obj 0 0 clip $1 $2 > > > > > > and it becomes this in Pd: > > > [clip 0 0] > > > > > > There's also sys_decodeddialog/sys_encodedialog that is used for > > > canvas_find. > > > > > > Additionally, I think you may have added another parser inside s_main > > > to separate filenames in Pd-l2ork (although that one doesn't have > anything > > > to do with dollarsigns). > > > > > > I have not yet decided where I want to add my own parser for the GUI > port. > > > I think a Pratt parser for message-box math would be neat. > > > > > > -Jonathan > > > > > > > > > > > > On Sunday, November 1, 2015 9:38 PM, Ivica Bukvic <[email protected]> wrote: > > > > > > > > > I presume this may pertain primarily to IOhannes and Miller, > > > > > > I am trying to figure out what is the purpose of raute2dollar in iemgui > > > objects? On an obvious level the function replaces # with an $ in a > send, > > > receive, and label symbols. Why is this necessary? Personally, I cannot > > > think of a reason why it would be, but then again, I may very well be > > > missing something... > > > > > > Best, > > > > > > Ico > > > > > > _______________________________________________ > > > [email protected] mailing list > > > UNSUBSCRIBE and account-management -> > > > http://lists.puredata.info/listinfo/pd-list > > > > > > > > > > > > _______________________________________________ > > [email protected] mailing list > > UNSUBSCRIBE and account-management -> > http://lists.puredata.info/listinfo/pd-list > >
_______________________________________________ [email protected] mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
