From what I can read in the documentation and the code, dot is only searched in a few hardcoded places and in the GRAPHVIZ_DOT variable. IMHO, it should be searched in that order:

• GRAPHVIZ_DOT
• let the system find it (just run dot)
• search for it in select places (actually, I don't think it makes sense to do that on Linux)

I'm running into this issue on Windows in the following use-case:

• I have dot in the PATH (but not in c:/program files)
• I'm using plantuml through doxygen
• In my build script, I don't want to hardcode the path to dot

The website http://plantuml.com/ seems to render with different languages based on browser hints, without the user having the possibility to change it. This is extremely annoying.

Please let the default language to English for everybody, and add links to the other languages (considering the it has many pages and not everything is translated completely)

Thank you.

For example, when I created an empty diagram, it will show:

Is there a way to remove the Welcome message?

Thank you :)

I updated plantuml to version 2017.08 and with new plantuml my tables are rendered incorrectly.

I have the following code:

@startuml
state X
note right of X #FFF
|= header 1 |= header 2 |= header 3 |
| //A// | abc | def |
| //B// | qwe | |
end note
@enduml

I expect to have this kind of render (like it used to be):

But the last version renders tables incorrectly:
(it's a ascreenshot)

It is clearly a regression since the version I used before. Unfortunately I don't know what version I was using.

Hello,

The library viz.js, which is a build of Graphviz in JavaCcript, supports multiple Graphviz engines such as circo, dot, neato, osage, or twopi.

I know in PlantUML, I can define dot language like below:

@startdot 
... my code 
@enddot

So is there a way to use other Graphviz engines other than dot?

Thank you :)

As I commented in the Any plan to move source code to github thread, I would love to help out with moving the project over from SourceForge to GitHub:

Is there anything I can do to help moving the active development of PlantUML to GitHub? I'm maintaining and contributing to a number of open source projects there and have been using Git professionally for 5 years, so I would love to contribute and help with the move. Having the source code and development process (issues, discussions, etc.) on SourceForge makes it a lot less visible and harder for people to contribute, I think.

Please let me know if there's anything I can do to assist. I can help out with answering questions, guiding the process to fit Git and GitHub's model, understanding the deal with pull requests, help plan branch strategies, etc. I have great interest in being able to create pull requests myself, so please let me know how I can help

Thanks!

@startuml
skinparam classFontSize 16
skinparam classStereotypeFontSize 16
skinparam classArrowFontSize 15
skinparam classAttributeFontSize 15
skinparam packageFontSize 16
package jizu {
  together {
    class JiZuYunXingZhuangTai << (E, orchid) >>
    class GuZhangLeiXing << (E, orchid) >>
    class JiZuRangeLeiXing << (E, orchid) >>
    class JiZuBianZaiLeiXing << (E, orchid) >>
  }
  together {
    class JiZu<J>
    interface JiZuCtrl
  }
  together {
    class JiZuCommonJi
    class ChaiYouJiJi
    class QiLunJiJi
    class AnDianJiJi
    class JiJi << (E, orchid) >>
  }
  interface JiZuCtrl <|- class JiZu
  interface AttrSetter <|- class JiZuCommonJi
  interface AttrSetter <|- class ChaiYouJiJi
  interface AttrSetter <|- class QiLunJiJi
  interface AttrSetter <|- class AnDianJiJi
  interface AttrSetter <|- class JiJi

}

@enduml

the generated diagram has a boring problem. it is always take all the class into a single line. this layout will make the diagram too wide. why not realizing auto wrap the class into a new line? If it can be, how can i Use this feature？

Is it possible to define a font stack like in CSS where multiple font names are given, and the first one that matches an installed font is used?
For example, I might prefer Helvetica, but that is not available on my colleagues Windows or Linux computer.
I would expect something like this to work:

skinparam defaultFontName Helvetica, Arial, "Open Sans"

Is this not implemented or is there a special syntax?

What Unicode font do you use at http://www.plantuml.com/plantuml?

this diagram shows fine on your server, but on our installation is missing the 5-digit Unicode chars such as <U+1F552>.

Our setup:

• linux 3.5.0-28-generic, java 1.6.0_25-b06, default encoding: UTF-8
• plantuml 1.2017.16 (20170903), confluence plugin 6.8-SNAPSHOT by avono AG (approx 20170907), graphviz 2.38.9 (20140413)

I thought it's a problem with the default encoding but after setting it to UTF-8 we still have the problem.

Using participant padding causing the creation arrows to get padded too.

@startuml
skinparam ParticipantPadding 30

participant a
create b
a --> b : <<create>>

@enduml

Produces:

Trying to render my first diagram and i have hit this error and unsure how to resolve ?

openjdk 9-Ubuntu
OpenJDK Runtime Environment (build 9-Ubuntu+0-9b161-1)
OpenJDK 64-Bit Server VM (build 9-Ubuntu+0-9b161-1, mixed mode)

Exception in thread "main" java.lang.IllegalAccessError: class net.sourceforge.plantuml.png.PngIOMetadata (in unnamed module @0x23941fb4) cannot access class com.sun.imageio.plugins.png.PNGMetadata (in module java.desktop) because module java.desktop does not export com.sun.imageio.plugins.png to unnamed module @0x23941fb4
	at net.sourceforge.plantuml.png.PngIOMetadata.writeWithMetadata(PngIOMetadata.java:60)
	at net.sourceforge.plantuml.png.PngIO.write(PngIO.java:86)
	at net.sourceforge.plantuml.png.PngIO.write(PngIO.java:80)
	at net.sourceforge.plantuml.ugraphic.g2d.UGraphicG2d.writeImageTOBEMOVED(UGraphicG2d.java:219)
	at net.sourceforge.plantuml.ugraphic.ImageBuilder.writeImageInternal(ImageBuilder.java:250)
	at net.sourceforge.plantuml.ugraphic.ImageBuilder.writeImageTOBEMOVED(ImageBuilder.java:172)
	at net.sourceforge.plantuml.sequencediagram.graphic.SequenceDiagramFileMakerPuma2.createOne(SequenceDiagramFileMakerPuma2.java:234)
	at net.sourceforge.plantuml.sequencediagram.SequenceDiagram.exportDiagramInternal(SequenceDiagram.java:225)
	at net.sourceforge.plantuml.UmlDiagram.exportDiagramNow(UmlDiagram.java:236)
	at net.sourceforge.plantuml.AbstractPSystem.exportDiagram(AbstractPSystem.java:134)
	at net.sourceforge.plantuml.SourceStringReader.outputImage(SourceStringReader.java:154)
	at net.sourceforge.plantuml.Pipe.managePipe(Pipe.java:95)
	at net.sourceforge.plantuml.Run.managePipe(Run.java:285)
	at net.sourceforge.plantuml.Run.main(Run.java:136)

IMO the association should be on a different line that the normal relation

I want to express:

1. a user is related to many routes and each route is related to one user
2. a UserReport combines a user, a route (and some extra fields)

@startuml

User "1" - "*" Route
UserReport ..(User, Route)

@enduml

the graphical presentation is confusing, a separate line should be used for the :

For example
I am writing a use-case diagram
a case named My Case, and it is a top-level menu,so I want it to be a larger size.

@startuml
bob -> alice : こんにちわ
bob <- alice : さようなら
@enduml

PNG works fine. But TUTXT, TXT doesn't work correctly. It seems platnum does not handle width of multi-byte characters. Below is output of java -jar platuml.jar -charset=UTF8 -tutxt input.platuml

https://gist.githubusercontent.com/mattn/d1be0d6711043abfce9946fa0112024a/raw/fe635b3a72724e75052c5c260d8e7ee7038b2a00/gistfile1.txt

Below is an expected output.

When defining diagrams with long texts on labels, the label texts are spreading the extent of the diagram. It is possible to manually enter \nin order to force a line break, but this adds syntactic clutter to the diagram source.

It would be great to be able to define a skinparam that allows for setting a maximum line width for label texts and have PlantUML inserting the line breaks automatically, e.g. MaximumLabelLineLength or similar.

Add Gradle as a build system since it's much easier to manage multi-module projects.

Thank you.

Please add support for ER Diagrams!
https://www.lucidchart.com/pages/er-diagrams

Thank you.

Hi,

please allow to add the the plantuml filename, inner page number and date of generation as a text to the generated images (like the various Fields in a Microsoft Word document.)

Thank you.

According the documentation the keyword newpage splits a big diagram into several pages. This only works with an empty diagram

@startuml

newpage

@enduml

This above code produces two empty pages, but produces an error when I add newpage between two classes:
Error line 3 in file: demo.puml Some diagram description contains errors

This code does not work right:

@startuml
state X {
  state Y
}
X: |= header 1 |= header 2 |= header 3 |
X: | //A// | abc | def |
X: | //B// | qwe | |
@enduml

At the same time, without inner state it works:

@startuml
state X {
}
X: |= header 1 |= header 2 |= header 3 |
X: | //A// | abc | def |
X: | //B// | qwe | |
@enduml

Could you please fix the first case?

the concept is nice, but it has one huge drawback: there is no way to more finely control diagram layouts.

result:

in other words, PlantUML is unusable in anything more than trivial.

I'm afraid this is more a rant than a problem, but please find a way to prevent such a spaghetti...

Hi,

Please add MindMap Diagram Support too :) .

Something compact like FreePlane https://sourceforge.net/projects/freeplane/ is producing: http://freeplane.sourceforge.net/wiki/index.php/File:FreeplaneWiki.jpg would be very useful.

Thanks in advance.

since sourceforge is down, there's no access to docs and even readme urls point to sourceforge

Hi there!

Don't you know, is there any tool to use PlantUML on iPad?
It would be great just to write diagram on iPad, save it to Dropbox and paste later into LaTeX-based documentation.

Before version 8025 I could have comments inside "skinparam" sections, as follows:

skinparam sequence {
    ArrowColor Black
    ActorBorderColor Black
    ActorBackgroundColor #BBBBBB
    LifeLineBorderColor Black
    LifeLineBackgroundColor #BBBBBB

    ParticipantBorderColor Black
    ParticipantBackgroundColor #AAAAAA

    BoxBorderColor Black

    'ParticipantFontName Impact
    'ParticipantFontSize 17
    'ParticipantFontColor #BBBBBB

    'ActorBackgroundColor aqua
    'ActorFontColor DeepSkyBlue
    'ActorFontSize 17
    'ActorFontName Aapex
}

With version 8025, I get a syntaxis error when doing this... which means that whatever change was added is not backwards compatible. Is this something that should be fixed or reverted back or should I change all of my files instead?

Hi,

How to have sub sequence auto number. For example 1, 2, and then 2.1, 2.2, 3, 4? Is this possible?

Thanks

When a task has multiple task dependencies the task should start after fulilling both dependencies.

For example in;

@startuml
[1] lasts 3 days
[2] lasts 8 days
[3] lasts 5 days

[3] starts at [2]'s end
[3] starts at [1]'s end
@enduml

[3] should start after [1] and [2] are finished. This is not actually working.

A temporal workaround is;

@startuml
[1] lasts 3 days
[2] lasts 8 days
[3] lasts 5 days

[3] starts at [1]'s end
[3] starts at [2]'s end
@enduml

@startuml

[*] --> S0
S0 : 初始状态

S0 --> [*] : EOF

S0 --> S1 : [
S1 --> S0 : ]
S1 : 分组状态

S0 --> S2 : ;#
S2 --> S0 : \\n
S2 : 行注释状态

S0 --> S3 : \\c
S3 --> S4 : =

S5 : 字符串状态
S4 --> S5 : "
S5 --> S4 : "

@enduml

the line S4 --> S5 : " dose not support escape "(quate mark), I suggest you can change your syntax parser as S4 --> S5 : "\"", S0 --> S1 : "hello, this is a comments string"

I've been using the Archimate feature solidly for a while now and would like to see it keep in line with the current specification.

The icon sets for Strategy and Physical layers are missing. Please find them below.

-graphvizdot no longer works in version 1.2017.12.

$ java -jar plantuml.jar -graphvizdot /opt/local/bin/dot example.puml
java.io.IOException: Cannot run program "/usr/bin/dot": error=2, No such file or directory
	at java.lang.ProcessBuilder.start(ProcessBuilder.java:1048)
	at java.lang.Runtime.exec(Runtime.java:620)
	at net.sourceforge.plantuml.cucadiagram.dot.ProcessRunner$MainThread.startThreads(ProcessRunner.java:163)
	at net.sourceforge.plantuml.cucadiagram.dot.ProcessRunner$MainThread.runJob(ProcessRunner.java:123)
	at net.sourceforge.plantuml.api.TimeoutExecutor$MyThread.run(TimeoutExecutor.java:79)
Caused by: java.io.IOException: error=2, No such file or directory
	at java.lang.UNIXProcess.forkAndExec(Native Method)
	at java.lang.UNIXProcess.<init>(UNIXProcess.java:247)
	at java.lang.ProcessImpl.start(ProcessImpl.java:134)
	at java.lang.ProcessBuilder.start(ProcessBuilder.java:1029)
	... 4 more
java.io.IOException: Cannot run program "/usr/bin/dot": error=2, No such file or directory
	at java.lang.ProcessBuilder.start(ProcessBuilder.java:1048)
	at java.lang.Runtime.exec(Runtime.java:620)
	at net.sourceforge.plantuml.cucadiagram.dot.ProcessRunner$MainThread.startThreads(ProcessRunner.java:163)
	at net.sourceforge.plantuml.cucadiagram.dot.ProcessRunner$MainThread.runJob(ProcessRunner.java:123)
	at net.sourceforge.plantuml.api.TimeoutExecutor$MyThread.run(TimeoutExecutor.java:79)
Caused by: java.io.IOException: error=2, No such file or directory
	at java.lang.UNIXProcess.forkAndExec(Native Method)
	at java.lang.UNIXProcess.<init>(UNIXProcess.java:247)
	at java.lang.ProcessImpl.start(ProcessImpl.java:134)
	at java.lang.ProcessBuilder.start(ProcessBuilder.java:1029)
	... 4 more

The environment variable GRAPHVIZ_DOT seems to work though.

$ GRAPHVIZ_DOT=/opt/local/bin/dot java -jar plantuml.jar example.puml

http://www.plantuml.com/plantuml/png/AqxCAU6gvghMud9Kg55mud86abfSjLor0000

Actual exception (will work once issue is fixed)

@startuml
salt
{
{*
A | B
C | *
}
}
@enduml

Hi,

Is there a way to disable the GUI?
Every time when I run plantuml from the command line, the starting image will pop up.

Is there a way to disable it?

Thanks for your fantastic work ;)

hide sterotype will cause most diagrams to hide the sterotype, but it doesn't in sequence diagrams.

• Link

@startuml
participant Bob <<Male>>
participant Alice <<Female>>
Bob -> Alice : hello

'this does nothing
hide stereotype
@enduml

I'm writing a plantuml parser and lexer for .net. Is there a master file or default keywords overview, BNF or AST?

Whilst a nice thought, the so called 'eggs' module is entirely unnecessary and extremely unwanted.

Today, any blank/placeholder UML diagrams are instead showing up as the "memorial" for the Paris attacks, as per here: https://github.com/plantuml/plantuml/blob/master/src/net/sourceforge/plantuml/eggs/PSystemMemorial.java

This is very unprofessional.

Hi All,

just noticed that the license on PlantUML is GPL3.

This is no problem as long as you are only using the libs/jar, but becomes a problem, if you want to integrate with it (e.g. building tools, using the API).

I think this will hinder adoption, not foster it.

A lot of companies have policies in place that do not allow to consume copyleft opensource projects.

Would suggest to change the license to something more permissive (BSD, MIT, Apache-2).

Regards ...
Roland

Not sure if this is the right place for a bug report, as it's about your website rather than the tool, but the list example on the creole webpage doesn't correctly render the sub list:

http://plantuml.com/creole

Good

$ plantuml -v
(0.000 - 123 Mo) 121 Mo - PlantUML Version 8031
(0.085 - 123 Mo) 119 Mo - GraphicsEnvironment.isHeadless() false

Every thing looks fine:

@startuml

node A{
    node B
    node C
}
node D

C <- B
B -> D

@enduml

Bad

Until I ad a label to the connection between B and D

@startuml

node A{
    node B
    node C
}
node D

C <- B
B -> D: E

@enduml

I want to add links to archimate nodes using syntax [[http://www.google.com]]

I receive a syntax error. I know that archimate support is added recently, perhaps this is not supported yet.

I woould like t o see it supported.

Please add a file to the root of the project, which states the licence under which this project is distributed.

Thank you very much in advance.

With this code:

!definelong TEST(name)
  !define JOIN(a,b,c) ##a##b##c
  JOIN(A,name,C)
!enddefinelong
TEST(B)

I get AnameC, but I want it to be ABC. How can I fix this?

I think it would be much simpler to parsing and validating plantUML files if there was a .g4 file containing the grammar for them (unless I have simply been unable to find one).
This has quite a few advantages:

1. Defining exactly which options and syntactic structure are valid is clear and simple.
2. Validating syntactically correct input files is trivial, since a Lexer and Parser can be built from the grammar file. This is nice also to create unit tests for builder-style structures.
3. Building listeners to handle the effects of parsing source files comes out of the box and are simple patterns.

Since my comment doesn't seem to have drawn any attention, but mostly because managing issues as comments isn't that good an idea anyway:

Hyperlink generation does not work for nested packages. Here's an example:

package org.jgrapes {
    package org.jgrapes.core [[org/jgrapes/core/package-summary.html#package.description]] {
    }

    package org.jgrapes.io {
    }

    package org.jgrapes.http {
    }
}

plantuml is an amazing tool and @arnaudroques provides amazing speed of innovation and a great support to developers and requests.

However, plantuml documentation is just sad: extremely outdated and fragmented across several places. I consider myself an advanced user (eg wrote a rdfpuml generator of diagrams from RDF data), and still I'm sadly surprised that I don't know half of its features. Just today I learned of

• Archimate diagrams
• ERD diagrams
• Gantt diagrams
• Tooltips in class diagrams

I mostly use http://plantuml.com/class-diagram, and it doesn't describe all features.
Another complication is that many features are described redundantly, eg Tooltips should be described in one place with a matrix stating where they are supported.

It's obvious to me that plantuml will flourish (people are asking for Venn, MindMaps, etc etc).
But when something is not documented, it's used by perhaps 1% of the potential users.
Furthermore, plantuml has a peculiar syntax, and if it's not meticulously documented for every feature, people can't use that feature.

So I'm willing to put a lot of time to make proper plantuml documentation, and help @arnaudroques and other contributors keep it up to date.

Sources (ADD MORE if you know others):

• https://github.com/arnaudroques/plantumlguide: latex, 7 years old, quite comprehensive (I can provide a PDF)
• java -jar plantuml.jar -language: lists all keywords.
  Supposed to be completely up to date, but eg PlantUML version 8019 (02 Feb 2015) was missing "skinparam shadowing", "skinparam packageStyle".
• pages at http://plantuml.com (a bit hard to read due to all the ads)
  • Sequence: Basic examples, Declaring participant, Use non-letters in participants, Message to Self, Change arrow style, Change arrow color, Message sequence numbering, Splitting diagrams, Grouping message, Notes on messages, Some other notes, Changing notes shape, Creole and HTML, Divider, Reference, Delay, Space, Lifeline Activation and Destruction, Participant creation, Incoming and outgoing messages, Stereotypes and Spots, Participants encompass, Removing Footer, Skinparam
  • Use Case: Usecases, Actors, Usecases description, Basic example, Extension, Using notes, Stereotypes, Changing arrows direction, Splitting diagrams, Left to right direction, Skinparam, Complete example
  • Class: Relations between classes, Label on relations, Adding methods, Defining visibility, Abstract and Static, Advanced class body, Notes and stereotypes, More on notes, Note on links, Abstract class and interface, Using non-letters, Hide attributes, methods..., Hide classes, Use generics, Specific Spot, Packages, Packages style, Namespaces, Automatic namespace creation, Lollipop interface, Changing arrows direction, Association classes, Skinparam, Skinned Stereotypes, Color gradient, Help on layout, Splitting large files,
  • Activity: Simple Activity, Label on arrows, Changing arrow direction, Branches, More on Branches, Synchronization, Long activity description, Notes, Partition, Skinparam, Octagon, Complete example
  • Activity BETA: Simple Activity, Start/Stop, Conditional, Repeat loop, While loop, Parallel processing, Notes, Colors, Arrows, Grouping, Swimlanes, Detach, SDL, Complete example
    Should be merged to the one above
  • Component: Components, Interfaces, Basic example, Using notes, Grouping Components, Arrows direction, Use UML2 notation, Long description, Individual colors, Sprite in Stereotype, Skinparam
  • State: Simple State, Composite state, Long name, Concurrent state, Arrow direction, Note, More in notes, Skinparam
  • Object: Definition of objects, Relations between objects, Adding fields, Visibility, Defines notes, Use packages, Skin the output, Split the image
  • Deployment: Declaring element, Linking, Packages, Round corner
  • Timing
• extra pages:
  • incubation: mish-mash of new features, should be dispatched to individual pages
  • skinparam: describes params and gives the default values, but is not complete
    A couple years ago there were 342 skinparams.

I think the best production method is Markdown (though I prefer Org-mode).
@arnaudroques, how do you make the web pages?

Preprocessor does not seem include external files after a newpage.

@startuml
!include x <- this works
newpage
!include x <- this fails silently
@enduml

I am using the Pycharm PlantUML integration plugin (version 1.9.0, using PlantUml library 8036)

It appears that the -computeurl/-encodeurl arguments are not honored when using the -pipe argument.

I read through some of the source code, and it looks like -pipe triggers a code path that completely ignores the -computeurl option.

Can you make -pipe honor -computeurl? If you can't do that, can you at least update the help content to say that they don't work together?

Hi,
A problem we are seeing is that when we have plantuml diagrams on a Confluence page multiple image attachments are generated whenever someone views the page.
This is resulting in anyone watching the page receiving many notifications that there has been an update, one for each image.
And this is happening for everyone who views the page.
The watchers do want to watch the page for edits but don't want to be notified by email 7 times when someone views it.

Is this a known issue and can it be disabled in some way?

Thanks a lot
Andreas

Many PlantUML tools uses PlantUML SVG export. Since ditaa diagrams can only be exported as PNG, many tools can not use the uselful ditaa syntax, such as the Atom markdown-preview-enhanced
plugin.

Setting the font skinparams for usecase diagrams seems to have no effect. The skinparams do seem to be listed in the language.

@startuml

skinparam {
  UsecaseStereotypeFontSize 20
  UsecaseStereotypeFontColor Red
  UsecaseStereotypeFontStyle bold
}

usecase (foo) <<bar>>
@enduml

As you can see, the stereotype is not big, red, and bold.