View previous topic :: View next topic |
Author |
Message |
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7931 Location: Salford, UK
|
Posted: Mon Sep 18, 2017 7:38 am Post subject: New DLLs |
|
|
A new set of Silverfrost DLLs is available for beta testing via the following link. This is mainly for the new native %pl winio@ format code.
Please make sure that you save backup versions of your existing DLLs before installing the new set.
https://www.dropbox.com/s/9e5veknlzz0snuw/newDLLs13.zip?dl=0 |
|
Back to top |
|
|
DanRRight
Joined: 10 Mar 2008 Posts: 2826 Location: South Pole, Antarctica
|
Posted: Mon Sep 18, 2017 10:29 am Post subject: |
|
|
Possibly wrong (older) DLLs? |
|
Back to top |
|
|
mecej4
Joined: 31 Oct 2006 Posts: 1888
|
Posted: Mon Sep 18, 2017 12:19 pm Post subject: |
|
|
No, these appear to be new. All the files in the Zip except salflibc64.lib have 2017-09-18 as the date.
There is a defunct link to an older Zip with the same name (newDLLS13.zip) in http://forums.silverfrost.com/viewtopic.php?t=3357, but clicking on that link gives you a 404 popup from Dropbox. Perhaps Paul can edit that post (29 Oct 2016) to indicate that the link is no longer valid.
It is not easy to find the latest newDLLSnn.zip using the forum search feature or using search engines such as Google, Yahoo, etc. Often, one does not know what value of "nn" to use, and there is no easy way of finding out where one can find the latest set. There is no corresponding problem with SDBG64.ZIP, but that may be so because only the latest release of SDBG64 is provided for sdbg64.exe. |
|
Back to top |
|
|
DanRRight
Joined: 10 Mar 2008 Posts: 2826 Location: South Pole, Antarctica
|
Posted: Mon Sep 18, 2017 12:44 pm Post subject: |
|
|
I see no differences |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7931 Location: Salford, UK
|
Posted: Mon Sep 18, 2017 1:19 pm Post subject: |
|
|
Dan
I guess that you mean that you see no differences in functionality.
I have fixed a few %pl bugs and added %pl[fixed_aspect]. You should at least be able to note that using %pl[fixed_aspect] does not raise an error condition. |
|
Back to top |
|
|
DanRRight
Joined: 10 Mar 2008 Posts: 2826 Location: South Pole, Antarctica
|
Posted: Mon Sep 18, 2017 9:14 pm Post subject: |
|
|
The %pl[fixed_aspect] keeps X and Y axis sizes of same ratio and does not cause crashes. Hopefully you figured out how to add scalability for all other elements of the plot. This will be main design shift from mentally sick braindead Simpleplot. Hope after that you will quickly fix all problems mentioned above. Having added few more 2003/2008 features like block/end block will add Silverfrost 1-2 new pro programmers, 99.9999% other will even not notice them. Even more, 99% do not need even Fortran95, happy with Fortran77. Do i need them? May be yes - may be not, the more i use Fortran to more i prefer simplicity, even primitivism. But making killing plotting facility will add thousands mostly novices but all pros will also appreciate it very much. This is like in economics: it's the cheap stuff for masses Walmart which is the #1 not the Nordstrom or Macy's which serve rich.
Much more valuably time will be spent on developing software testing robot which will automatically verify new builds.
Centered X axis caption works OK. But new DLLs completely broke another LOG example (one mentioned in that thread about new PL, no external data needed, with the red and blue symbols)
One more design thing: i do not like very much the new naming winop@ introduced, and now both winio@ and winop@ are used, which causes nothing but confusion. Is it possible to make for new %pl that it will use only winio@ or winop@ ? Even better name exists for it - winpl@ |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7931 Location: Salford, UK
|
Posted: Tue Sep 19, 2017 8:05 am Post subject: |
|
|
Dan
I confess that I struggle with your belligerence. However, if you will reproduce the sample that now fails then I will aim to take a look at it.
The winop@ routine is provided to make things easier. If you don't like it then you don't need to use it. In fact, the same principle applies to the whole. |
|
Back to top |
|
|
DanRRight
Joined: 10 Mar 2008 Posts: 2826 Location: South Pole, Antarctica
|
Posted: Tue Sep 19, 2017 11:41 am Post subject: |
|
|
Sorry, for the style. Admittedly, I know this Eastern style indeed looks really bad on the West and trying to avoid it. It goes from the argumentative style of the entire scientific community in my home country created in his time by the famous but very rude Nobel Prize physicist Landau. Pity that not many people use new %pl to articulate those several pages of bugs and wishes Western way. In short all those bugs are still there.
/*
Most striking difference of these styles I first time have experienced on some meeting many years ago when I just came to the West when one guy was defending his poor weak half-cooked PhD. I was like on fire on my chair ready to destroy him but was silent due to my very poor English at that time. All ordeal went shockingly amicably for me and the only objection came from the chair of meeting who also clearly understand that the PhD is total BS with the absurd claims contradicting all known data, conservation laws and common sense. He just asked politely: "Are you happy with this your result?". For me this question was like a cultural shock.
On the other half of the world Landau and everyone in similar situation would go on full voice trying to totally bury the person: "You have really really surprised all us here. All your work is total idiotism and nonsense. Are you in your right mind? You can not read or count 2x2? "
To avoid public humiliation of basically everyone, in such cases Landau usually heard the person before his talk one-to-one and told him what he thinks about him and his work: "OK, I will allow this garbage to be heard by the people in this seminar but with so much @#$%^ absurd in this work I am now very low opinion about your abilities so you better take the broom and shovel and never do the science again"
Now I hope you understand what was in my head seeing and not believing that pages and pages of found bugs were not fixed, even the most disastrously looking and even few new were made.
Last edited by DanRRight on Tue Sep 19, 2017 6:37 pm; edited 2 times in total |
|
Back to top |
|
|
John-Silver
Joined: 30 Jul 2013 Posts: 1520 Location: Aerospace Valley
|
Posted: Tue Sep 19, 2017 12:14 pm Post subject: |
|
|
mecej4 ..... I learned the hard way DLLs revert to 1 after a new release arrives, so since we had 8.1 arrive earlier this year the numbering reverted.
Maybe DLLs need numbering DDLLs_<ver>_nn e.g in current case DLLs_8.1_13.
I see Paul has followed the suggestion and added a note to the relevant comment on the post you quoted ...... but not for th ex2 other DLLs that appear there !!!!
I suggested quite a qhile back that new DLLs should be gathered somewhere specifically (as well as on the specific related post) so that people who don't follow all posts are aware easily of any 'new arrivals'. I think I suggested a new Forum Topic would be a good option. However it seems there are problems creating new forum topics, I've suggested them for other things like documentation etc .. .too, so I guess something else would be required.
It is clear however by this case that confusion on the menu of the day, especially for 'casual' users. |
|
Back to top |
|
|
John-Silver
Joined: 30 Jul 2013 Posts: 1520 Location: Aerospace Valley
|
Posted: Tue Sep 19, 2017 12:22 pm Post subject: |
|
|
Paul, I think it's unfair, way OTT to refer to Dan's comments as belligerent when in fact they just come I think from frustration.
. |
|
Back to top |
|
|
John-Silver
Joined: 30 Jul 2013 Posts: 1520 Location: Aerospace Valley
|
Posted: Tue Sep 19, 2017 2:59 pm Post subject: |
|
|
In updating to the dll13 in the ftn95 installation directory I nor�ticed the presence of the following files:
clearwin64.a
clearwin64f.a
clearwin64f.dll
What exactly are these files and why do they not also need updating ? |
|
Back to top |
|
|
mecej4
Joined: 31 Oct 2006 Posts: 1888
|
Posted: Tue Sep 19, 2017 3:29 pm Post subject: Re: |
|
|
DanRRight wrote: | ... famous but very rude Nobel Prize physicist Landau. Pity that not many people use new %pl to articulate those several pages of bugs and wishes Western way. |
I remember reading a multi-letter exchange between E.W.Dijkstra and J.W.Backus (both Turing Award winners) in such tones, and an explanation elsewhere that this open, often rude, style was common among Dutch Academics. See, for example, https://github.com/jiahao/backus-dijkstra-letters-1979/blob/master/pdf/all-letters.pdf . Do Web search for "Backus Dijkstra letters" and "This guy�s arrogance takes your breath away".
[In Jest]Perhaps we could add to the FTN95 wish list a request for a %bL control item to provide for Dan's needs. The bL could stand for belligerent Landau" or "bug List", take your choice. [/Jest] |
|
Back to top |
|
|
PaulLaidler Site Admin
Joined: 21 Feb 2005 Posts: 7931 Location: Salford, UK
|
Posted: Tue Sep 26, 2017 9:42 am Post subject: |
|
|
Dan
Regarding your comment "But new DLLs completely broke another LOG example", I have found a sample program that matches the description but it runs OK for me. This may be a result the latest fixes (not yet released) or I might have modified your %pl styles to get it to work.
The simplest way forward is for you to try the next version when it comes out. |
|
Back to top |
|
|
John-Silver
Joined: 30 Jul 2013 Posts: 1520 Location: Aerospace Valley
|
Posted: Mon Oct 02, 2017 12:46 am Post subject: |
|
|
Dan, I presume the 'broken' example is the one here ...
http://forums.silverfrost.com/viewtopic.php?t=3487&postdays=0&postorder=asc&start=30
comment headed 'CONTINUED 2')
I ran it on my machne, 32bit, with dlls13
As you say the xaxis titles are now centred.
y axs ttles are centred IF you change to a linear scale, but for log-linear , and after removing your 'title offset adjustments' it's not centred on axis, it appears to be centred over length of yaxis + height of title. They still interfere with axis tick labels though, so offsets will always be required
however. for linear scalesthe lines are NOT PLOTTED just pts/symbols!!!
However, for Log-Linear , cycling through with the button, I saw several times on the last 1, 2, or 3 line segments OVERSHOOTS of the line segments beyond the symbols !.
I did a one-off experiment putting the x-axis range at -30 to 60 and the results are lets day not the best, even with the default linear-linear axes scales.
In fact I got a lot of cont�stant flickering and an incomplete window for say 30 seconds, then the window appears but any attempt to resize it, or after simply waiting doing nothing and the program crashes out.
So my conclusion is - title centring is almost ok for linear-linear and a little less almost ok for log-linear. However, there are stil lstrange things afoot with both log AND linear scales with this particular example |
|
Back to top |
|
|
DanRRight
Joined: 10 Mar 2008 Posts: 2826 Location: South Pole, Antarctica
|
Posted: Tue Oct 03, 2017 2:40 pm Post subject: |
|
|
John, Would be nice if you show the pictures, so I will compare with mine and also this will help Paul to check if these problems were the ones which as he wrote are already fixed |
|
Back to top |
|
|
|